Skip to main content

Table 3 Mean precision (P), recall (R), and F1-measure (F) 10-fold cross validation results on training data for the most frequent properties per scene

From: Personality-dependent content selection in natural language generation systems

 

Majority class baseline

Personality-dependent CS

Scene

Property

Inst.+/ −

P

R

F

P

R

F

01

red-sandals

46/105

0.70

1.00

0.82

0.82

0.96

0.88

01

brown-dress

37/114

0.76

1.00

0.86

0.86

0.99

0.92

01

long-dress

26/125

0.83

1.00

0.91

0.84

0.98

0.90

02

black-canvas

74/77

0.51

1.00

0.68

0.81

0.99

0.89

02

white-bucket

65/86

0.57

1.00

0.73

0.76

0.98

0.86

02

large-bucket

30/121

0.80

1.00

0.89

0.86

1.00

0.92

03

fat-lips

89/62

0.56

1.00

0.71

0.77

0.99

0.87

03

white-bedspread

59/92

0.62

1.00

0.76

0.74

0.99

0.85

03

red-details

56/95

0.54

1.00

0.78

0.75

0.99

0.85

04

black-person

75/76

0.52

1.00

0.69

0.84

0.97

0.90

04

raised-arms

35/116

0.62

1.00

0.76

0.73

1.00

0.84

05

sitting-people

38/113

0.75

1.00

0.86

0.86

0.96

0.81

05

green-coat

38/113

0.75

1.00

0.86

0.77

1.00

0.87

05

white-hair

28/123

0.81

1.00

0.90

0.77

1.00

0.87

06

pink-bedspread

79/72

0.52

1.00

0.68

0.79

0.99

0.88

06

blue-blanket

74/77

0.51

1.00

0.68

0.83

0.97

0.89

06

blue-coat

67/84

0.56

1.00

0.71

0.77

0.99

0.87

08

black-man

111/40

0.72

1.00

0.84

0.86

0.96

0.91

08

blue-cap

35/116

0.77

1.00

0.87

0.72

1.00

0.84

08

barbed-wire

25/126

0.83

1.00

0.91

0.85

1.00

0.92

09

yellow-shorts

51/100

0.54

1.00

0.70

0.75

0.98

0.85

09

plaid-shorts

45/106

0.70

1.00

0.82

0.80

0.99

0.89

09

blue-shirt

37/114

0.75

1.00

0.86

0.77

0.97

0.86

10

dry-grassland

40/111

0.74

1.00

0.85

0.83

0.99

0.90

10

elderly-woman

30/121

0.80

1.00

0.89

0.83

0.99

0.90

10

red-pants

30/121

0.80

1.00

0.89

0.75

1.00

0.86

 

Mean

51/100

0.68

1.00

0.80

0.80

0.99

0.88

  1. Best F1 results are highlighted